Citation: Cell Death and Disease (2017) 8, e2866; doi:10.1038/cddis.2017.257 OPEN Official journal of the Cell Death Differentiation Association www.nature.com/cddis ARRDC4 regulates enterovirus 71-induced innate immune response by promoting K63 polyubiquitination of MDA5 through TRIM65

Jun Meng1,2,6, Zhenyu Yao3,6, Yaqing He1, Renli Zhang1, Yanwei Zhang1, Xiangjie Yao1, Hong Yang1, Long Chen1, Zhen Zhang1, Hailong Zhang1, Xueqin Bao4, Gang Hu5, Tangchun Wu*,2 and Jinquan Cheng*,1

Enterovirus 71 (EV71) is the main causative agent of hand, foot and mouth disease (HFMD), which induces significantly elevated levels of cytokines and chemokines, leading to local or system inflammation and severe complications, whereas the underlying regulatory mechanisms and the inflammatory pathogenesis remain elusive. ARRDC4 is one member of family, having important roles in glucose metabolism and G--coupled receptors (GPCRs) related physiological and pathological processes, however, the function of ARRDC4 in innate immune system is largely unknown. Here we identified that ARRDC4 expression was increased after EV71 infection in THP-1-derived macrophages and verified in EV71-infected HFMD patients and the healthy candidates. The expression level of ARRDC4 was positively correlated with the serum concentration of IL-6, TNF-α and CCL3 in clinical specimens. ARRDC4 interacted with MDA5 via the -like N domain, and further recruited TRIM65 to enhance the K63 ubiquitination of MDA5, resulting in activation of the downstream innate signaling pathway and transcription of proinflammatory cytokines during EV71 infection. Our data highlight new function of ARRDC4 in innate immunity, contributing to the better understanding about regulation of MDA5 activation after EV71 infection, and also suggest ARRDC4 may serve as a potential target for intervention of EV71-induced inflammatory response. Cell Death and Disease (2017) 8, e2866; doi:10.1038/cddis.2017.257; published online 8 June 2017

Enterovirus 71 (EV71) is a single positive-stranded RNA virus detect long fragments viral RNA (44kb) mainly from picorna- belonging to the picornaviridae family. EV71 is the main virus and paramyxovirus.8–10 Potentiated by E3 ligase-mediated causative agent of hand, foot and mouth disease (HFMD), K63 polyubiquitination, RIG-I and MDA5 are activated and which induces a wide range of clinical manifestations leading to subsequently activate the essential adaptor protein mitochon- severe neurological complications in infants and children drial antiviral signaling (MAVS). MAVS further recruits down- younger than 5 years old.1 EV71-infected patients carry high stream adaptors to activated transcription factors IFN regulatory level of cytokines and chemokines,2,3 whereas exhibit impaired factors (IRFs), NF-κB and activator protein 1 (AP-1), inducing production of type I interferon (IFN).4,5 In the absence of type I production of proinflammatory cytokines and type I IFNs.11 IFNs, the other cytokines and chemokines are irreplaceable to RLR-triggered innate immune signaling is tightly regulated defense against EV71 invasion, however, the disordered to achieve effective viral clearance without immunopathology. production of inflammation mediators may lead to cell or tissue Polyubiquitination is one of the key regulatory mechanisms for injury. The detailed regulation mechanism and the inflammatory RLRs. Upon recognition of viral RNA, RIG-I undergoes K63 pathogenesis after EV71 infection remain elusive. polyubiquitination catalyzed by tripartite motif protein 25 or Once viruses invade the host, the innate immune system Riplet to activate innate signaling pathway,12,13 whereas ring- detects viral nucleic acid with pattern recognition receptors finger protein 125 (RNF125) acts as negative regulators and (PRRs), including Toll-like receptors (TLRs), retinoic acid- mediates K48-linked polyubiquitination of RIG-I and MDA5, inducible -I (RIG-I)-like receptors (RLRs) and other leading to their degradation through proteasome.14 The cytosolic sensors.6 The crucial members of RLRs, RIG-I and regulation of RIG-I ubiquitination has been extensively – melanoma differentiation-associated gene 5 (MDA5), have non- investigated,12 18 however, the regulation mechanism of redundant roles in cytosolic viral RNA sensing.7 MDA5 prefers to MDA5 ubiquitination is poorly understood.

1Department of Microbiology, Shenzhen Center for Disease Control and Prevention, Shenzhen, Guangdong, China; 2Department of Occupational and Environmental Health Key Laboratory of Environment and Health, Ministry of Education and State Key Laboratory of Environmental Health (Incubating), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China; 3Department of Translational Medicine R&D Center, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, China; 4Department of Epidemiology and Biostatistics, College of Public Health, Zhengzhou University, Zhengzhou, Henan, China and 5Department of Biostatistics and Epidemiology, College of Public Health, Sun Yat-SenUniversity, Guangzhou, Guangdong, China *Corresponding author: T Wu, Key Laboratory of Environment and Health, Ministry of Education & Ministry of Environmental Protection, and State Key Laboratory of Environmental Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China. Tel: +86 27 8369 2333; Fax: +86 27 8369 2560; E-mail: [email protected] or J Cheng, Shenzhen Center for Disease Control and Prevention, Shenzhen, Guangdong, China. Tel: +86 755 2562 6488; Fax: +86 755 2553 2595; E-mail: [email protected] 6These authors contributed equally to this work. Received 17.1.17; revised 16.4.17; accepted 04.5.17; Edited by A Oberst ARRDC4 promotes EV71-induced MDA5 activation JMenget al 2

Figure 1 EV71 infection upregulates ARRDC4 expression. (a) Microarray analysis of EV71-infected t-MØ (MOI = 50) for indicated time. Heat map represents arrestins family expression change (Po0.05). (b) Q-PCR analysis of expression of α-arrestin members in t-MØ treated with EV71 for 24 h or untreated. (c) mRNA expression of ARRDC4 was assayed by Q-PCR in t-MØ infected with EV71 for indicated time. (d) Immunoblot analysis of ARRDC4, EV71-VP1 and β-actin expression in lysates of t-MØ stimulated with EV71 for shown time. Data are shown as mean ± S.D. of triplicate samples (b and c), or are representative of three independent experiments with similar results (d)

Arrestins family consists of visual arrestins (S antigen (SAG) Results and arrestin 3 (ARR3)), β-arrestin 1–2 (ARRB1 and ARRB2), and α-arrestins (arrestin domain-containing 1–5 (ARRDC1–5) ARRDC4 expression is upregulated in response to EV71 and thioredoxin-interacting protein (TXNIP)), which have infection. In order to identify molecules selectively involved essential and versatile roles in the signaling of G-protein- in the regulation of EV71-triggered innate immune response, coupled receptors (GPCRs).19 In recent years, some mem- we screened differential expression molecules in EV71- bers of the arrestins family have been reported involving in infected THP-1-derived macrophages (t-MØ) through Gene- more diversiform physical and pathological processes. Much Chip Microarrays, and identified all the members of arrestins of the concerns are about the β-arrestins, especially in family were up or downregulated (Figure 1a). ARRDC4 was immune and inflammatory responses.20 β-Arrestin can nega- the most significantly upregulated after EV71 infection tively regulate LPS triggered proinflammatory cytokines (Figure 1a). We validated the expression of α-arrestins 21 production, and have moderates roles in antiviral innate subfamily members in EV71-infected t-MØ by Q-PCR, and immune response.22 ARRDC4 is one member of α-arrestins, got the same results as the microarrays (Figure 1b). More- which has the arrestin-like structure and highly conserved over, the mRNA and protein levels of ARRDC4 were polyproline (PY) motif in the C-terminal tail.19 ARRDC4 increased in response to EV71 infection in a time- participates in glucose metabolism,23 and exhibits more varied dependent manner (Figures 1c and d). These results indicate functions through recruiting some E3 ligases to induce that ARRDC4 may be involved in EV71-induced immune ubiquitination of GPCRs.24 Interestingly, it has been sug- response. gested that α-arrestin and β-arrestin may hetero-associate and have coordinated or antagonistic functions depending on ARRDC4 promotes EV71-triggered proinflammatory context,25 however, the function of ARRDC4 in immune cytokines production. To identify the role of ARRDC4 in response is largely unknown, thus deserves fuller exploration. EV71-triggered immune responses, we silenced ARRDC4 Herein, we analyzed profiles in EV71- with specific siRNA in t-MØ, and detected significantly infected macrophages with GeneChip Microarrays, notably, all knockdown of ARRDC4 expression in both mRNA and of the arrestins family members were up or down regulated, protein levels (Figures 2a and b). ARRDC4 silencing and ARRDC4 exhibited the highest elevated level. In EV71- antagonized the mRNA expression and the production of infected clinical specimens, the expression level of ARDDC4 IL-6, TNF-α and CCL3 in t-MØ infected with EV71 (Figures 2c was positively correlated with the serum concentration of and d). However, type I IFN (IFN-α and IFN-β) expression in proinflammatory cytokines. ARRDC4 interacted with MDA5 mRNA level was inhibited severely by EV71 as reported and promoted K63 ubiquitination of MDA5 via TRIM65, before,5 and knockdown of ARRDC4 did not affect type I IFN consequently enhanced MDA5-dependent signaling path- production (data not shown). We further observed the ways, leading to strengthening of cytokines production in antiviral effects of ARRDC4, and found that ARRDC4 anti-EV71 innate immune response. silencing promoted EV71 replication in t-MØ (Figure 2e),

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation J Meng et al 3

Figure 2 ARRDC4 promotes proinflammatory cytokines production in response to EV71 infection. (a and b) Q-PCR and immunoblot analysis of ARRDC4 knockdown efficiency in t-MØ 48 h after transfected with control siRNA or ARRDC4 siRNA. (c) Q-PCR analysis of IL-6, TNF-α and CCL3 mRNA expression in t-MØ, which were transfected as in (a) and infected with EV71 for 12 h. (d) ELISA of cytokines in supernatants of t-MØ treated as in (c). (e) Q-PCR analysis of EV71-VP1 RNA expression in EV71-infected t-MØ transfected as in (a). (f) Tissue culture infective dose (TCID50) assay in supernatants of t-MØ treated as indicated. Data are shown as mean ± S.D. of triplicate samples, *Po0.05 (a, c-f), or are representative of three independent experiments with similar results (b) accordingly, the viral titer in the supernatant was increased as with control subjects (Figure 3a). We evaluated the correla- well (Figure 2f). These data suggest that ARRDC4 positively tion between ARRDC4 expression and the serum concentra- regulates anti-EV71 innate immune response through tion of proinflammatory cytokines in patients, and found that promoting IL-6, TNF-α and CCL3 production. ARRDC4 mRNA expression level was positively correlated with the concentration of IL-6 (r = 0.4806, Po0.01), TNF-α (r = 0.6685, Po0.001) and CCL3 (r = 0.5998, Po0.001) ARRDC4 expression is positively correlated with cytokines (Figure 3b). These data further verify that ARRDC4 could production in EV71-infected HFMD patients. We confirmed ARRDC4 expression in peripheral blood mononuclear cells positively regulate EV71-induced innate proinflammatory (PBMCs) derived from 30 EV71-infected HFMD children and cytokines production in HFMD patients. 30 healthy candidates, which were age (HFMD 38.23 ± 2.880 months, healthy control 40.67 ± 2.156 months, ARRDC4 enhances MDA5-triggered signaling activation P = 0.5015) and sex (HFMD 60.0% male, healthy control upon EV71 infection. We further investigated the molecular 56.7% male, P = 0.7934) matching. The expression levels of mechanism through which ARRDC4 augmented innate ARRDC4 were much higher in patients’ PBMCs compared proinflammatory cytokines production. Upon EV71 infection,

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation JMenget al 4

Figure 3 ARRDC4 expression is positively correlated with serum proinflammatory cytokines concentration in EV71-infected HFMD patients and healthy controls. (a) Q-PCR analysis of ARRDC4 mRNA expression in PBMCs from EV71-infected HFMD patients (n = 30) and controls (n = 30), which normalized by the mean expression value. Data are shown as mean ± S.D., **Po0.01. (b) The correlation between ARRDC4 mRNA expression level (assayed by Q-PCR) and IL-6, TNF-α, CCL3 serum concentration (assayed by ELISA) in patients were evaluated with the Pearson correlation test

ARRDC4 silenced t-MØ exhibited impaired phosphorylation of MAPKs ERK, JNK, p38 and NF-κB subunit p65, and less degradation of IκBα, accordingly, the protein level of EV71- VP1 was increased in ARRDC4 silenced t-MØ (Figure 4). But the activation of IRF3 was weak and similar in both ARRDC4 knockdown and control t-MØ stimulated with EV71 (Figure 4), which was consistent with the impaired IFN production.4,5 These data demonstrate that ARRDC4 promotes activation of NF-κB and MAPK signaling pathways upon EV71 infection. The PRRs that recognize EV71 and trigger the activation of innate immune signaling are still elusive. MDA5, which is critical for picornavirus detection, has been reported have important roles during EV71 infection.26 In EV71-infected t-MØ, MDA5 silencing significantly inhibited the activation of ERK, JNK, p38, p65, IκBα and elevated the protein level of EV71-VP1 (Supplementary Figure S1a and c). On the other hand, the PRRs TLR3 and TLR7/8, which can recognize dsRNA or ssRNA derived from virus, may also participate in EV71-triggered innate immune signaling. Then we knockdown the expression of TLR8 in t-MØ (Supplementary Figure S1b), and observed TLR8 silencing inhibited the activation of the aforementioned signaling to some extent after EV71 infection, however, the effect was too weakly to promote the expression of EV71-VP1 (Supplementary Figure S1d). Owing to the expression of TLR3 were undetectable in mRNA and protein level in THP-1 cells (data not shown), which was 27 κ consistent with previous report, we overexpressed NF- B Figure 4 ARRDC4 enhances EV71-triggered activation of NF-κB and MAPK luciferase reporter plasmid with TRIF and ARRDC4 in pathway. The t-MØ were transfected with ARRDC4-specific siRNA or the negative HEK293T cells, and found ARRDC4 did not influence TRIF- control siRNA. After 48 h, the cells were infected with EV71 for indicated time, then induced activation of NF-κB luciferase reporter (Supplem- total and phosphor-JNK, -ERK, -p38, -p65, -IRF3 and total IκBα were detected with β entary Figure S1e). These data indicate that, during EV71 immunoblot, the expression of EV71-VP1 and -actin were detected as well. Data are representative of three independent experiments with similar results infection, MDA5 is the most indispensable PRRs for the activation of innate signaling pathway, whereas the other TLRs may partially have some assistant roles. Collectively, the data tandem mass spectrometry assay, we identified ARRDC4 above suggest that ARRDC4 promotes EV71-induced inflam- could interacted with MDA5. Then, we corroborated endo- matory cytokines production through positively regulation of genous ARRDC4 could interact with MDA5 in t-M in MDA5-triggered innate signaling activation. Ø reciprocal immunoprecipitation experiments (Figures 5a and b). Moreover, we confirmed the interaction between the ARRDC4 interacts with MDA5 via the arrestin-like N overexpressed ARRDC4 and MDA5 in HEK293T cells domain. Given the positive function of ARRDC4 in regulat- (Figures 5c and d). To determine the domains of ARRDC4 ing EV71-induced innate immune signaling, we sought to for interaction with MDA5, we constructed various ARRDC4 determine the potential target of ARRDC4. Using immuno- truncation mutants, and found the full-length ARRDC4 or precipitation of endogenous ARRDC4 in EV71-infected t-MØ ARRDC4 with only the arrestin-like N domain (ARRDC4-N) plus reverse-phase nanospray liquid chromatography- could be co-immunoprecipitated with MDA5 (Figure 5e). The

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation J Meng et al 5

Figure 5 ARRDC4 interacts with MDA5 via arrestin-like N domain. (a and b) The t-MØ were infected with EV71, immunoprecipitation (IP) with anti-ARRDC4 (a) or anti-MDA5 (b) and control IgG antibodies. Immnoblot analysis of the interaction between ARRDC4 and MDA5; input, the whole-cell lysates without immunoprecipitated. (c and d) HEK293T cells were transfected with Flag- or Myc-tagged ARRDC4 and MDA5 plasmids, after 48 h, IP with anti-Flag antibodies, then detected protein interaction by immunoblot. (e) Schematic structure of ARRDC4 and the derivatives were shown, full-length ARRDC4 (ARRDC4-WT), arrestin-like N domain (ARRDC4-N), arrestin-like C domain (ARRDC4- C), PPXY motif (ARRDC4-PPXY). The IP and immunoblot analysis of the interaction between MDA5 and the indicated ARRDC4 constructs in HEK293T cells. Data are representative of three independent experiments with similar results findings indicate that ARRDC4 interacts with MDA5 via the ligase, which was associated with ARRDC4 to promote K63 ARRDC4-N domain after EV71 infection. ubiquitination of MDA5 upon EV71 infection. It has been reported recently that TRIM65 specifically interacts with ARRDC4 is required for K63 ubiquitination of MDA5 upon MDA5 and facilitates K63 ubiquitination of MDA5 at lysine EV71 infection. Next, we wondered how ARRDC4 regulated 743, leading to oligomerization and activation of MDA5 after innate immune signaling pathway through MDA5. MDA5 encephalomyocarditis virus (EMCV) infection.28 Thus, we senses viral RNA and was activated by K63 polyubiquitin further observed whether there was cooperation between chains mediating by E3 ubiquitin ligases, leading to down- TRIM65 and ARRDC4 in promoting the activation of MDA5. stream signal transduction.11 As ARRDC4 contains highly Through immunoprecipitation, we corroborated endogenous conserved PY motifs, which can recruit E3 ubiquitin ligases ARRDC4 could interact with TRIM65 and MDA5 in EV71- 24 resulting in ubiquitination of target molecular, we consid- infected t-MØ (Figure 7a) and in HEK293T cells (data not ered whether ARRDC4 influenced K63 ubiquitination of shown). The interaction between overexpressed ARRDC4 MDA5. In response to EV71 infection, MDA5 could be K63 and TRIM65 were further confirmed in HEK293T cells ubiquitinated, which was significantly impaired in ARRDC4 (Figures 7b and c). ARRDC4 silencing could reduce the knockdown t-MØ (Figure 6a). In addition, we found that interaction between TRIM65 and MDA5 in EV71-infected overexpression of ARRDC4 augmented K63 ubiquitination of t-MØ (Figure 7d), suggesting that ARRDC4 may acts as the MDA5 after EV71 infection, but had no effect on K0 (mutant adaptor for the interaction of TRIM65 and MDA5. Further- ubiquitin with no lysine residues retained) ubiquitination of more, we observed that ARRDC4 only elevated low level of MDA5 (Figure 6b). The ARRDC4 mutant lacking ARRDC4-N K63 ubiquitination of MDA5 without EV71 infection, whereas (ARRDC4-ΔN), which could not interact with MDA5, was this could significantly be raised up when ARRDC4 were co- unable to enhance K63 ubiquitination of MDA5 (Figure 6c). transfected with TRIM65 (Figure 7e), indicating TRIM65 was These data suggest that ARRDC4 interacts with MDA5 and required for ARRDC4 to modulate MDA5 activation. In promotes K63 ubiquitination activation of MDA5 leading to addition, TRIM65-catalyzed K63 ubiquitination of MDA5 was innate signaling cascade during EV71 infection. augmented by ARRDC4 (Figure 7e). ARRDC4 facilitated TRIM65- and MDA5-induced activation of NF-κB luciferase ARRDC4 promotes MDA5 K63 ubiquitination through reporter, however, the positive regulatory effect of ARRDC4 TRIM65. Finally, we would like to search for the E3 ubiquitin was disappeared without the ARRDC4-N domain (Figure 7f).

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation JMenget al 6

Figure 6 ARRDC4 mediates K63 ubiquitination of MDA5. (a) Immunoblot analysis of K63 polyubiquitination of the precipitated MDA5 in t-MØ, which were transfected and stimulated as indicated. (b and c) HEK293T cells were transfected with indicated plasmids and infected with EV71 for 12 h, followed by IP with anti-Flag antibody. Immunoblot analysis of ubiquitination of MDA5; Ub-K0, mutant ubiquitin with no lysine residues retained. Data are representative of three independent experiments with similar results

Figure 7 ARRDC4 promotes MDA5 K63 ubiquitination through TRIM65. (a) The t-MØ were infected with EV71, and IP with anti-ARRDC4 and control antibodies, followed by immunoblot analysis as indicated. (b and c) IP and immunoblot analysis of protein interaction in HEK293T cells transfected with Flag- or Myc-tagged ARRDC4 and TRIM65. (d) Immunoblot analysis of the association between precipitated TRIM65 and MDA5, in t-MØ, which were transfected with ARRDC4 siRNA and control siRNA for 48 h and stimulated with EV71 for indicated time. (e) IP and immnunoblot analysis of ubiquitination of MDA5 in HEK293T cells transfected with indicated plasmids. (f) HEK293T cells were transfected with NF-κB luciferase reporter plasmid together with the other indicated plasmids, after 48 h, the luciferase activity was determined. Mock means empty vector. Data are representative of three independent experiments with similar results (a-e), or are shown as mean ± S.D. of triplicate samples, *Po0.05, **Po0.01 (f)

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation J Meng et al 7

These data indicate that, upon EV71 infection, ARRDC4 important for oligomerization and activation of MDA5,31 interacts with MDA5 and further recruits TRIM65 to mediate however, the regulatory mechanisms are still incomplete K63 polyubiquitin chains on MDA5, leading to downstream known. TRIM13 interacts with MDA5 and negatively regulates signaling pathways activation. MDA5-mediated innate signaling upon EMCV infection,32 nevertheless, whether and how TRIM13 mediates ubiquitina- Discussion tion modification of MDA5 is still elusive. RNF123 inhibits IFN- β production through competitively inhibiting MAVS binding to Upon EV71 infection, host innate immune system sets MDA5/RIG-I, however, the negative regulatory function of antivirus strategies to produce proinflammatory cytokines RNF123 is independent on its E3 ligase activity.33 Recently, and chemokines to restrain and clear the pathogens, mean- TRIM65 was found specifically interacts with MDA5 and while, virus can also utilizes multiple factors to neutralize and facilitates K63 ubiquitin chains on MDA5 at lysine 743, leading escape innate immune surveillance. The underlying mechan- 28 isms of host–virus interaction during EV71 infection attract to oligomerization and activation of MDA5. Herein, we found increasing concerns. We have shown here that ARRDC4 TRIM65 was involved in EV71-induced K63 ubiquitination of interacted with MDA5 and positively regulated the K63 MDA5, further confirmed the non-redundant roles of TRIM65 polyubiquitination of MDA5 through TRIM65, leading to in MDA5 activation. In addition, ARRDC4 acts as the adaptor activation of the downstream innate immune signaling and for promoting the interaction between MDA5 and TRIM65, and production of cytokines, thus contributed to defense EV71 TRIM65 is required for ARRDC4 to modulate MDA5 activation. infection. Our study also confirmed the positive correlation We provide a new regulation manner of MDA5 activation, between ARRDC4 expression level and cytokines production which ARRDC4 and TRIM65 collaboratively enhance K63 in clinical specimens, suggesting that ARRDC4 is increased to ubiquitination of MDA5, further work will be put forward to enhance anti-EV71 innate immunity. Some severe EV71- explore the underlying mechanism. infected HFMD patients, who carried extremely high level of During the virus life cycle, the EV71 non-structural proteins proinflammatory cytokines, were accompanied with much 2A and 3C have essential roles in host–virus interactions.34 In higher expression of ARRDC4, indicating that the aberrant serum and cerebrospinal fluid of EV71-infected HFMD increase of ARRDC4 may related to the imbalanced release of patients, the production of type I IFNs are impaired.4,5 cytokines and the pathological injury during EV71 infection. Accumulating evidence indicate that the 2A and 3C protein We considered that ARRDC4 may serve as a potential target can target some crucial molecules in innate immune signaling for moderating dysregulation of EV71-induced inflammation. pathways to block type I IFN production. In EV71 2A- ARRDC4 is one of α-arrestins family members, containing transfected Hela cells, the 2A protein can cleave MDA5 and arrestin-like N domain, arrestin-like C domain and highly MAVS, leading to repression of type I IFN transcription.35 The conserved polyproline (PY) motif in the C-terminal tail. It’s activation and nuclear translocation of IRF3 were extremely known that ARRDC4 participates in glucose metabolism 35,36 through inhibiting glucose uptake with its arrestin-like inhibited by 2A and 3C proteins. In the EV71-infected domains.19,23 In our research, the upregulated ARRDC4 t-MØ, we also found that the phosphorylation of IRF3 was interacted with MDA5 and recruited E3 ubiquitin ligase suppressed and the production of type I IFN were undetect- TRIM65 for MDA5 activation, having a new function in EV71- able. In addition, EV71 can also target type I IFN-signaling, the triggered innate immune response. We identified that the 2A protein induces IFNAR1 degradation,5 and the 3C protein arrestin-like N domain of ARRDC4 was required for the cleaves IRF737 and IRF9.38 In our study, we found that MDA5 interaction with MDA5. Our data confirmed that the arrestin- are only partially cleaved in EV71-infected t-MØ. We speculate like domains of arrestins family may have critical roles in that, once EV71 recognized by MDA5, the activation of TBK1- mediating interaction with other molecules. The PY motif of IRF3 signaling pathway was extremely impaired as mentioned ARRDC4 is as important as the arrestin-like domains, above, however, the remained MDA5, which was not cleaved especially for interaction with WW domain-containing proteins, by EV71 viral proteins, could significantly activate innate 24 which include several ubiquitin ligases such as Nedd4. signaling pathway, leading to much higher level of cytokines β β ARRDC4 binds to 2 adrenergic receptor ( 2AR) and recruits production. The activation of MDA5 was promoted by Nedd4 with the PY motif to mediate ubiquitination and ARRDC4 and TRIM65, otherwise, there could be more trafficking of β AR.29 Whether ARRDC4 recruits TRIM65 2 regulation mechanisms involved in anti-EV71 innate immunity, through the PY motif or the other domains is still need to be and crosstalk between different pathways and different uncovered. Moreover, we still wonder the mechanisms molecular need to be further clarified. underlying ARRDC4 upregulation. It is known that the Collectively, we have shown that ARRDC4 interacts with expression of ARRDC4 is upregulated by transcription factor MDA5 and promotes K63 polyubiquitination of MDA5 via MondoA and Mlx complex resulting in restricting glucose uptake and cell growth.30 We estimate that, upon EV71 TRIM65, consequently activates innate signaling pathway, infection, some transcript activators would bind to the leading to enhanced anti-EV71 immune response. Our promoter of ARRDC4 to enhance transcription of ARRDC4, findings suggest a new function of ARRDC4 in innate immune and the detail mechanism will be carried out in our further system, and provide new insight into the regulatory mechan- research. ism of EV71-induced inflammation, which line out a direction MDA5 has momentous roles in innate immune responses for intervention imbalanced inflammatory response during against RNA virus.9,10 The K63 polyubiquitination of MDA5 is EV71 infection.

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation JMenget al 8

Materials and Methods RNA quantification. The total cellular RNA was extracted with Trizol reagent Cell culture and virus. THP-1 and HEK293T cell lines were obtained from (Invitrogen Corporation), and the cDNA synthesis and Quantitative real-time PCR American Type Culture Collection (Manassas, VA, USA). The base medium for THP-1 were performed as described previously.39 The primers used for ARRDC4 were is RPMI-1640 medium (Gibco, Waltham, MA, USA) with fetal bovine serum (FBS) 5′-GCAGGAAAGAGTCGCCCG-3′ (sense), 5′-TAAACTTGTGCGGCAATCCTG-3′ (Gibco, Grand Island, NY, USA) to a final concentration of 10%. The base medium for (antisense). The others specific primers for Q-PCR assay were, IL-6 forward HEK293T is DMEM (Gibco, Waltham, MA, USA) with 10% FBS (Gibco, Grand Island, 5′-AAATTCGGTACATCCTCGACGG-3′, IL-6 reverse 5′-GGAAGGTTCAGGTT NY, USA). THP-1 cells were seeded into cell plates and were incubated with PMA GTTTTCTGC-3′; TNF-α forward 5′-ATGAGCACTGAAAGCATGATCC-3′, TNF-α (50nM) for 24 h, the differentiated THP-1-derived macrophages got attached to the reverse 5′-GAGGGCTGATTAGAGAGAGGTC-3′; CCL3 forward 5′-AGTTCTCTG plate bottom, then were washed with fresh media for use. The EV71/SZ50/CHN/2014 CATCACTTGCTG-3′, CCL3 reverse 5′-CGGCTTCGCTTGGTTAGGAA-3′. β-Actin isolate (subgenotype C4a, GenBank: KT428649) used in this study was from the was used as internal quantitative control. department of microbiology, Shenzhen Center for Disease Control and Prevention (CDC). Virus was amplified in RD cells and indentified via RT-PCR with EV71 specific Immunoprecipitation and immunoblot analysis. Total proteins of cells primers. Viral titer was determined by TCID50 on RD cells. were prepared using cell lysis buffer and protease inhibitor mixture, then quantified the concentrations with the bicinchoninic acid assay BCA (Pierce, Rockford, IL, Microarray analysis. THP-1-derived macrophages were infected with EV71 USA). Equal amount of proteins were used for immunoprecipitation and immunoblot for indicated time, and the total RNA was isolated using Trizol reagent (Invitrogen analysis as described.39 For Flag-tag protein immunoprecipitation, anti-Flag M2 Corporation, Carlsbad, CA, USA), and the RNA was validated with Agilent Array affinity gel (Sigma) was used and performed follow the technical procedure. platform. The microarray analysis was performed by Gminix Biotechnology Company (Shanghai, China) with Affymetrix Human Transcriptome Array 2.0 Luciferase reporter assay. HEK293T cells were co-transfected with (HTA 2.0) (Affymetrix, Waltham, MA, USA). Each transcript was represented by a luciferase reporter plasmid, pRL-TK-Renilla luciferase plasmid and other constructs specific exon or splice junction probe, which can identify individual transcripts as indicated in the context. Total amounts of plasmid DNA were equalized with precisely. empty control vector. Luciferase activities were measured with the Dual Luciferase Reporter Assay System (Promega, Madison, WI, USA) according to the Study design and populations. Followed the principle of case–control manufacturer’s instructions. Data are normalized for transfection efficiency by study, we recruited 30 HFMD patients and 30 healthy controls in Shenzhen dividing firefly luciferase activity with activity of Renilla luciferase. Children’s Hospital, Shenzhen Baoan District People’s Hospital and Shajing Institution of Disease Prevention and Healthcare from 2015 to 2016. All of the Statistical analysis. The statistical significance of comparisons between two patients were diagnosed with HFMD according to the WHO Guide to Clinical groups was determined with two-tailed Student’s t-test. Categorical characteristics Management and Public Health Response for HFMD, and were further confirmed by were examined by the χ2 test. Correlation between ARRDC4 expression level and EV71 isolation and identification in clinical samples, such as stool, rectal and throat cytokines production level in clinical specimens were analyzed using Spearman swabs. Patients who were treated with drugs after infection were excluded. Healthy correlation test. All statistical analysis performed using SPSS 17.0 (Chicago, IL, candidates were from physical health examination children, who were matched by USA). P-values of o0.05 were considered statistically significant. age, sex with the patients, and diagnosed without any infection disease and further etiological confirmed without EV71 infection. The PBMCs were isolated from the blood samples within 2 h by Ficoll paque premium from GE Healthcare Life Science Conflict of Interest (Marlborough, MA, USA) and stored in Trizol reagent at –801C. All of the The authors declare no conflict of interest. participants were informed consent and the study was approved by the Ethics Committee of the Shenzhen Center for Disease Control and Prevention. Acknowledgements. We thank Dr. Xingguang Liu and Dr. Chaofeng Han for helpful discussion; and Dr. Rongbin Zhou (University of Science and Technology of Reagents and antibodies. PMA were purchased from Sigma (St. Louis, MO, China, Hefei, China) for providing the TRIM65 constructs. We thank Dr. Kai Zhao and USA). Antibodies to ARRDC4 were from Abcam Inc. (ab209679) (Cambridge, MA, Dr. Suli Huang for technical assistance; and Gminix company (Shanghai, China) for USA) or Santa Cruz Biotechnology Inc. (sc-135444) (Dallas, TX, USA). Antibodies bioinformatics support. This work was supported by Grants from the National Natural to MDA5 (#21775-1-AP) was from Proteintech Group (Rosemont, IL, USA). Science Foundation of China (81601386 and 81371815) and Science and Antibodies to MDA5 (D74E4, #5321), ERK (#9102), p-ERK (#9101), JNK (56G8, Technology Research Projects of Shenzhen (JCYJ20160428142411948 and #9258), p-JNK (81E11, #4668 ), p38 (#9212), p-p38 (#9211), p65 (D14E12, #8242), JCYJ20150401145529020), and the National Major Science and Technology p-p65 (93H1, #3033), IκBα (L35A5, #4814), IRF3 (D83B9, #4302) and p-IRF3 Research Projects of China (2016YFC1202001). (#4947) were from Cell Signaling Technology (Beverly, MA, USA). Antibody to TRIM65 (HPA021578) was from Sigma, anti-EV71-VP1 antibody (ab169442) was from Abcam Inc. The antibodies to Flag-tag (#14793), Myc-tag (#2276), HA-tag Author contributions (#3724) and V5-tag (D3H8Q, #13202) were from Cell Signaling Technology. The JC and YH designed and supervised the study; JM, ZZ and XB collected the clinical anti-K63-ubiquitin (#05-1313) was purchased from Millipore (San Diego, CA, USA). samples; JM, XY, HY, LC and HZ isolated and validated virus from the clinical samples; JM, ZY, YH, RZ, YZ and GH conducted the study; JM, ZY, JC and TW Plasmids and transfection. Recombinant vectors encoding full-length and analyzed the data and wrote the paper. mutant human ARRDC4 (GenBank No. NM_183376.2) and IFIH1 (for MDA5) (GenBank No. NM_022168.3) were cloned into pcDNA3.1 vectors (Invitrogen Corporation) as described previously.39 The Flag-tagged TRIM65 constructs was 1. Solomon T, Lewthwaite P, Perera D, Cardosa MJ, McMinn P, Ooi MH. Virology, provided by Dr. Rongbin Zhou (University of Science and Technology of China, epidemiology, pathogenesis, and control of enterovirus 71. Lancet Infect Dis 2010; 10: 778–790. Hefei, China). All constructs were confirmed by sequencing. Plasmids were 2. Wang SM, Lei HY, Yu CK, Wang JR, Su IJ, Liu CC. Acute chemokine response in the blood transiently transfected into HEK293T cells with jetPEI reagents (Polyplus and cerebrospinal fluid of children with enterovirus 71-associated brainstem encephalitis. Transfection Company, Illkirch, France) following the manufacturer instructions. J Infect Dis 2008; 198: 1002–1006. 3. Khong WX, Foo DG, Trasti SL, Tan EL, Alonso S. Sustained high levels of interleukin-6 RNA interference. For transient knockdown of ARRDC4, the siRNA duplexes contribute to the pathogenesis of enterovirus 71 in a neonate mouse model. J Virol 2011; 85: – specific for ARRDC4 were transfected with the INTERERin according to the 3067 3076. 4. Liu ML, Lee YP, Wang YF, Lei HY, Liu CC, Wang SM et al. Type I interferons protect mice manufacturer protocol (Polyplus Transfection Company). The sequence of siRNA – ′ ′ against enterovirus 71 infection. J Gen Virol 2005; 86: 3263 3269. specific for human ARRDC4 was 5 -GAGAAGCUAUUCCAAUCUAUU-3 . The 5. Lu J, Yi L, Zhao J, Yu J, Chen Y, Lin MC et al. Enterovirus 71 disrupts interferon signaling by ′ ′ negative control siRNA sequence was 5 -UUCUCCGAACGUGUCACGUUU-3 . reducing the level of interferon receptor 1. J Virol 2012; 86: 3767–3776. These siRNA were designed and synthesized by GenePharma Co. (Shanghai, 6. Cao X. Self-regulation and cross-regulation of pattern-recognition receptor signalling in China). health and disease. Nat Rev Immunol 2016; 16:35–50.

Cell Death and Disease ARRDC4 promotes EV71-induced MDA5 activation J Meng et al 9

7. Yoneyama M, Kikuchi M, Natsukawa T, Shinobu N, Imaizumi T, Miyagishi M et al. The RNA 27. Carpenter S, Wochal P, Dunne A, O'Neill LA. Toll-like receptor 3 (TLR3) signaling helicase RIG-I has an essential function in double-stranded RNA-induced innate antiviral requires TLR4 interactor with leucine-rich REPeats (TRIL). J Biol Chem 2011; 286: responses. Nat Immunol 2004; 5: 730–737. 38795–38804. 8. Kato H, Takeuchi O, Mikamo-Satoh E, Hirai R, Kawai T, Matsushita K et al. Length- 28. Lang X, Tang T, Jin T, Ding C, Zhou R, Jiang W. TRIM65-catalized ubiquitination is essential dependent recognition of double-stranded ribonucleic acids by retinoic acid-inducible gene-I for MDA5-mediated antiviral innate immunity. J Exp Med 2017; 214: 459–473. and melanoma differentiation-associated gene 5. J Exp Med 2008; 205: 1601–1610. 29. Han SO, Kommaddi RP, Shenoy SK. Distinct roles for beta-arrestin2 and arrestin-domain- 9. Gitlin L, Barchet W, Gilfillan S, Cella M, Beutler B, Flavell RA et al. Essential role of mda-5 in containing proteins in beta2 adrenergic receptor trafficking. EMBO Rep 2013; 14: 164–171. type I IFN responses to polyriboinosinic:polyribocytidylic acid and encephalomyocarditis 30. Kaadige MR, Looper RE, Kamalanaadhan S, Ayer DE. Glutamine-dependent anapleurosis picornavirus. Proc Natl Acad Sci USA 2006; 103: 8459–8464. dictates glucose uptake and cell growth by regulating MondoA transcriptional activity. 10. Kato H, Takeuchi O, Sato S, Yoneyama M, Yamamoto M, Matsui K et al. Differential roles of Proc Natl Acad Sci USA 2009; 106: 14878–14883. MDA5 and RIG-I helicases in the recognition of RNA viruses. Nature 2006; 441: 101–105. 31. Jiang X, Kinch LN, Brautigam CA, Chen X, Du F, Grishin NV et al. Ubiquitin-induced 11. Wu J, Chen ZJ. Innate immune sensing and signaling of cytosolic nucleic acids. Annu Rev oligomerization of the RNA sensors RIG-I and MDA5 activates antiviral innate immune Immunol 2014; 32: 461–488. response. Immunity 2012; 36: 959–973. 12. Gack MU, Shin YC, Joo CH, Urano T, Liang C, Sun L et al. TRIM25 RING-finger E3 ubiquitin 32. Narayan K, Waggoner L, Pham ST, Hendricks GL, Waggoner SN, Conlon J et al. TRIM13 is ligase is essential for RIG-I-mediated antiviral activity. Nature 2007; 446: 916–920. a negative regulator of MDA5-mediated type I interferon production. J Virol 2014; 88: 13. Oshiumi H, Miyashita M, Inoue N, Okabe M, Matsumoto M, Seya T. The ubiquitin ligase 10748–10757. Riplet is essential for RIG-I-dependent innate immune responses to RNA virus infection. Cell 33. Wang S, Yang YK, Chen T, Zhang H, Yang WW, Song SS et al. RNF123 has an E3 ligase- Host Microbe 2010; 8: 496–509. independent function in RIG-I-like receptor-mediated antiviral signaling. EMBO Rep 2016; 14. Arimoto K, Takahashi H, Hishiki T, Konishi H, Fujita T, Shimotohno K. Negative regulation of 17: 1155–1168. the RIG-I signaling by the ubiquitin ligase RNF125. Proc Natl Acad Sci USA 2007; 104: 34. Tsou YL, Lin YW, Shao HY, Yu SL, Wu SR, Lin HY et al. Recombinant adeno-vaccine 7500–7505. expressing enterovirus 71-like particles against hand, foot, and mouth disease. PLoS Negl 15. Kuniyoshi K, Takeuchi O, Pandey S, Satoh T, Iwasaki H, Akira S et al. Pivotal role of Trop Dis 2015; 9: e0003692. RNA-binding E3 ubiquitin ligase MEX3C in RIG-I-mediated antiviral innate immunity. Proc 35. Feng Q, Langereis MA, Lork M, Nguyen M, Hato SV, Lanke K et al. Enterovirus 2Apro targets Natl Acad Sci USA 2014; 111: 5646–5651. MDA5 and MAVS in infected cells. J Virol 2014; 88: 3369–3378. 16. Friedman CS, O'Donnell MA, Legarda-Addison D, Ng A, Cardenas WB, Yount JS et al. The 36. Lei X, Liu X, Ma Y, Sun Z, Yang Y, Jin Q et al. The 3C protein of enterovirus 71 inhibits tumour suppressor CYLD is a negative regulator of RIG-I-mediated antiviral response. retinoid acid-inducible gene I-mediated interferon regulatory factor 3 activation and type I EMBO Rep 2008; 9: 930–936. interferon responses. J Virol 2010; 84: 8051–8061. 17. Chen W, Han C, Xie B, Hu X, Yu Q, Shi L et al. Induction of Siglec-G by RNA viruses inhibits 37. Lei X, Xiao X, Xue Q, Jin Q, He B, Wang J. Cleavage of interferon regulatory factor 7 by the innate immune response by promoting RIG-I degradation. Cell 2013; 152: 467–478. enterovirus 71 3C suppresses cellular responses. J Virol 2013; 87:1690–1698. 18. Cui J, Zhu L, Xia X, Wang HY, Legras X, Hong J et al. NLRC5 negatively regulates the 38. Hung HC, Wang HC, Shih SR, Teng IF, Tseng CP, Hsu JT. Synergistic NF-kappaB and type I interferon signaling pathways. Cell 2010; 141: 483–496. inhibition of enterovirus 71 replication by interferon and rupintrivir. J Infect Dis 2011; 203: 19. Alvarez CE. On the origins of arrestin and . BMC Evol Biol 2008; 8: 222. 1784–1790. 20. Jiang D, Xie T, Liang J, Noble PW. beta-Arrestins in the immune system. Prog Mol Biol Transl 39. Meng J, Liu X, Zhang P, Li D, Xu S, Zhou Q et al. Rb selectively inhibits innate IFN-beta Sci 2013; 118: 359–393. production by enhancing deacetylation of IFN-beta promoter through HDAC1 and HDAC8. 21. Wang Y, Tang Y, Teng L, Wu Y, Zhao X, Pei G. Association of beta-arrestin and TRAF6 J Autoimmun 2016; 73:42–53. negatively regulates Toll-like receptor-interleukin 1 receptor signaling. Nat Immunol 2006; 7: 139–147. Cell Death and Disease is an open-access journal 22. Mo W, Zhang L, Yang G, Zhai J, Hu Z, Chen Y et al. Nuclear beta-arrestin1 functions as a scaffold for the dephosphorylation of STAT1 and moderates the antiviral activity of published by Nature Publishing Group. This work is IFN-gamma. Mol Cell 2008; 31: 695–707. licensed under a Creative Commons Attribution 4.0 International 23. Patwari P, Chutkow WA, Cummings K, Verstraeten VL, Lammerding J, Schreiter ER et al. License. The images or other third party material in this article are Thioredoxin-independent regulation of metabolism by the alpha-arrestin proteins. J Biol ’ Chem 2009; 284: 24996–25003. included in the article s Creative Commons license, unless indicated 24. Becuwe M, Herrador A, Haguenauer-Tsapis R, Vincent O, Leon S. Ubiquitin-mediated otherwise in the credit line; if the material is not included under the regulation of endocytosis by proteins of the arrestin family. Biochem Res Int 2012; 2012: Creative Commons license, users will need to obtain permission from 242764. the license holder to reproduce the material. To view a copy of this 25. Shea FF, Rowell JL, Li Y, Chang TH, Alvarez CE. Mammalian alpha arrestins link activated seven transmembrane receptors to Nedd4 family e3 ubiquitin ligases and interact with beta license, visit http://creativecommons.org/licenses/by/4.0/ arrestins. PLoS ONE 2012; 7: e50557. 26. Lei X, Cui S, Zhao Z, Wang J. Etiology, pathogenesis, antivirals and vaccines of hand,foot, and mouth disease. Natl Sci Rev 2015; 2: 268–284. r The Author(s) 2017

Supplementary Information accompanies this paper on Cell Death and Disease website (http://www.nature.com/cddis)

Cell Death and Disease